Векторные представления изображений
otherentity
ID: imageembeddings
Related Papers
Sheffield MultiMT: Using Object Posterior Predictions for Multimodal Machine Translation
This paper describes the University of Sheffield’s submission to the WMT17 Multimodal Machine Translation shared task. We participated in Task 1 to de- velop an MT system to translate an …...
Understanding Spatial Relations through Multiple Modalities
AbstractRecognizing spatial relations and reasoning about them is essential in multiple applications including navigation, direction giving and human-computer interaction in general. Spatial relations between objects can either be explicit – …...